Performing appID based firewall services on a host

Abstract

Some embodiments of the invention provide a novel architecture for capturing contextual attributes on host computers that execute one or more machines, and for consuming the captured contextual attributes to perform services on the host computers. The machines are virtual machines (VMs) in some embodiments, containers in other embodiments, or a mix of VMs and containers in still other embodiments. Some embodiments execute a guest-introspection (GI) agent on each machine from which contextual attributes need to be captured. In addition to executing one or more machines on each host computer, these embodiments also execute a context engine and one or more attribute-based service engines on each host computer. One of these service engines is a firewall engine. Through the GI agents of the machines on a host, the context engine of that host in some embodiments collects contextual attributes associated with network events and/or process events on the machines. The context engine then provides the contextual attributes to the firewall engine, which, in turn, use these contextual attributes to identify firewall rules to enforce.

Description

BACKGROUND

Middlebox services have historically been hardware appliances that are implemented at one or more points in a network topology in an enterprise or a datacenter. With the advent of software defined networking (SDN) and network virtualization, traditional hardware appliances do not take advantage of the flexibility and control that is provided by SDN and network virtualization. Accordingly, in recent years, some have suggested various ways to provide middlebox services on hosts. Most of these middlebox solutions, however, do not take advantage of the rich contextual data that can be captured for each data message flow on the host. One reason for this is that existing techniques do not provide an efficient, distributed scheme for filtering the thousands of captured contextual attributes in order to efficiently process service rules that are defined in terms of much smaller sets of contextual attributes.

BRIEF SUMMARY

Some embodiments of the invention provide a novel architecture for capturing contextual attributes on host computers that execute one or more machines, and for consuming the captured contextual attributes to perform services on the host computers. The machines are virtual machines (VMs) in some embodiments, containers in other embodiments, or a mix of VMs and containers in still other embodiments.

Some embodiments execute a guest-introspection (GI) agent on each machine from which contextual attributes need to be captured. In addition to executing one or more machines on each host computer, these embodiments also execute a context engine and one or more attribute-based service engines on each host computer. Through the GI agents of the machines on a host, the context engine of that host in some embodiments collects contextual attributes associated with network events and/or process events on the machines. As further described below, the context engine then provides the contextual attributes to the service engines, which, in turn, use these contextual attributes to identify service rules that specify context-based services to perform on processes executing on the machines and/or data message flows sent by or received for the machines.

In some embodiments, the context engine of a host collects contextual attributes from the GI agents of the machines on that host through a variety of different ways. For instance, in some embodiments, the GI agent on a machine registers hooks (e.g., callbacks) with one or more modules (e.g., kernel-space modules or user-space modules) in the machine's operating system for all the new network connection events and all the new process events.

Upon occurrence of a new network connection event, the GI agent receives a callback from the OS and based on this callback, provides a network event identifier to the context engine. The network event identifier provides a set of attributes pertaining to the network event. These network event attributes in some embodiments include a five-tuple identifier (i.e., source port and IP address, destination port and IP address, and protocol) of the requested network connection, a process identifier of the process requesting the network connection, a user identifier associated with the requesting process, and a group identifier (e.g., an activity directory (AD) identifier) associated with the requesting process.

In some embodiments, the context engine directs the GI agent to collect from the OS modules additional process parameters, associated with the process identifier (ID) that it received with the network event. These additional process parameters in some embodiments include the process name, the process hash, the process path with command line parameters, the process network connection, the process-loaded modules, and one or more process consumption parameters specifying the process' consumption of one or more resources of the machine (e.g., central processing unit consumption, network consumption, and memory consumption). Instead of using the process identifier to query the GI agent for additional process parameters associated with a network event, the context engine in other embodiments receives all the process parameters associated with a network event in one shot, when the GI agent reports the network event to the context engine.

The OS on a machine in some embodiments holds up a new network event (i.e., does not start sending data messages for the network event) until the GI agent on the machine directs it to proceed with processing the network event. In some of these embodiments, the GI agent only allows the OS to proceed with processing the network event after the context engine has collected all the needed attributes for this event (e.g., after receiving a message from the context engine that specifies that it has received all the process or network attributes that it needs for the new network event).

In some embodiments, the context engine uses the process hash that it receives from the GI agent to identify the name and version of the application (i.e., the software product), to which the process belongs. To do this, the context engine in some embodiments stores process hashes and associated application names/versions, compares the process hash that it receives from the GI agent with the stored process hashes to identify a matching hash, and then uses the application name/version of the matching hash as the application name and version of the process associated with the event.

In some embodiments, the context engine obtains the process hashes and application names/versions from one or more network or compute managers, which may operate on another device or computer. In other embodiments, the context engine provides the hash associated with a process identifier to a network or compute manager, which then matches this hash to its process hash records and provides the application name/version of the associated process to the context engine. Once the context engine obtains the application name/version associated with a network event, it can provide the name and version attributes to the attribute-based service engines, which can use this information (i.e., the application name and/or version) to identify the service rule to enforce.

Upon occurrence of a process event, the GI agent receives a callback from the OS and based on this callback, provides a process event identifier to the context engine. The process event identifier provides a set of attributes pertaining to the process event. This set of attributes includes the process identifier in some embodiments. In some embodiments, this set also includes a user identifier and a group identifier (e.g., an activity directory (AD) identifier).

In some embodiments, the GI agent provides all the process parameters (e.g., process identifier, user ID, group ID, process name, process hash, loaded module identifiers, consumption parameters, etc.) associated with a process event to the context engine when it reports the process event to the context engine. In other embodiments, the context engine directs the GI agent to collect from the OS modules additional process parameters, associated with the process identifier that the context engine received with the process event. These additional process parameters in some embodiments are the same (e.g., process name, process hash, loaded module identifiers, consumption parameters, etc.) as the process parameters mentioned above for reported network events.

The context engine of some embodiments augments the contextual attributes that it receives from the GI agent with contextual attributes that it receives from other modules that execute on the host. For instance, in some embodiments, a deep packet inspection (DPI) module executes on the host. The context engine or another module (e.g., a firewall engine) directs this DPI engine to examine data messages of a data message flow associated with a process ID, to identify the type of traffic being sent in these data messages by the application associated with the process ID.

The identified traffic-type identity is today commonly referred to as the AppID. Also currently, there are a number of DPI modules that analyze the messages of a data message flow to generate the AppID. In some embodiments, the context engine combines the AppID that it obtains for a network event with other context attributes that it identifies for this event (e.g., by using the event's five-tuple identifier to associate the AppID with the collected contextual attributes), in order to produce a very rich set of attributes that the service engines can then use to perform their services. This rich set of attributes provides true application identity (i.e., the application name, application version, application traffic type, etc.), based on which the service engines can perform their services.

Also, in some embodiments, a threat detection module executes on the host computer along with the context engine. Once the context engine obtains a set of process parameters that specify that a process has started on a machine or that is sending data messages on the machine the context engine in some embodiments provides one or more process parameters (e.g., process hash, application name, application version, AppID, other process parameters, etc.) to the threat detection module. This threat detection module then generates a threat level indicator (e.g., low, medium, high, etc.) for the identified process and provides this threat level indicator to the context engine. The context engine then provides this threat level indicator to one or more service engines, as another contextual attribute for performing services on a new process event or the data messages of a new network event. A service engine can use the threat level indicator as another attribute to identify service rules to enforce.

The context engine employs a push model in some embodiments to distribute the collected contextual attributes to the service engines, employs a pull model in other embodiments to distribute these attribute to the service engines, while in other embodiments, employing a push model for some service engines and a pull model for other service engines. In the push model, the context engine distributes to a service engine the contextual attributes that it collects for a process event or a network event with the process' identifier and/or the network event's flow identifier (e.g., the flow's five-tuple identifier). In some embodiments, the context engine distributes to the service engine only the contextual attributes that are relevant for that service engine's service rules.

In the pull model, the context engine receives queries from a service engine for the contextual attributes that the context engine has collected for a particular process or network connection. In some embodiments, the context engine receives a process ID or a flow identifier (e.g., five-tuple identifier), with a query from the service engine, and uses the received identifier to identify the attribute set that it has to provide to the service engine. In some embodiments, the context engine generates a service token (also called a service tag) for the collection of attributes that are relevant for the service engine, and provides this service token to another module (e.g., the GI agent or another module on the host) to pass along to the service engine (e.g., pass along in a data message's encapsulating tunnel header). The service engine then extracts the service token and provides this service token to the context engine in order to identify the contextual attributes that the context engine has to provide to the service engine.

The context engine in some embodiments provides contextual-attributes to several context-based service engines on its host computer. In some embodiments, the context engine and the service engines are all kernel space components of a hypervisor, in which multiple VMs or containers execute. In other embodiments, the context engine and/or one or more service engines are user space processes. For example, one or more service engines in some embodiments are service VMs (SVMs).

Different embodiments use different types of context-based service engines. For instance, in some embodiments, the attribute-based service engines include (1) a firewall engine that performs context-based firewall operations on data messages sent by or received for the machines, (2) a process control engine that enforces context-based process control operations (e.g., process assessment and termination operations) on processes started on the machines, (3) a load balancing engine that performs context-based load balancing operations to distribute the data message flows from the machines to different destination or service nodes in one or more destination/service node clusters, and (4) an encryption engine that performs context-based encryption or decryption operations to encrypt data message from the machines, or to decrypt data messages received for the machines.

Another context-based service engine in some embodiments is a discovery service engine. In some embodiments, the discovery engine captures new process events and new network events from the context engine, along with the contextual attributes that the context engine collects for these process and network events. The discovery service engine then relays these events and their associated contextual attributes to one or more network managers (e.g., servers) that provide a management layer that allows network administrators to visualize events in a datacenter and specify policies for compute and network resources in the datacenter.

In relaying these events and attributes to the network management layer, the discovery module of some embodiments performs some pre-processing of these events and attributes. For example, in some embodiments, the discovery module filters some of the network or process events, while aggregating some or all of these events and their attributes. Also, in some embodiments, the discovery engine directs the context engine to collect additional contextual attributes for process or network events through the GI agents or other modules (e.g., the DPI engine or threat detection engine), or to capture other types of events, such as file events and system events.

For example, in some embodiments, the discovery engine directs the context engine to build an inventory of the applications installed on the machines, and to periodically refresh this inventory. The discovery engine might so direct the context engine at the request of the management plane, or based on operational configurations that the management or control plane specifies for the discovery engine. In response to the request from the discovery engine, the context engine in some embodiments has each GI agent on each of its host's machines discover all the installed processes on the machine, and all the running processes and services.

After building an inventory of installed applications and the running processes/services, the discovery engine of a host computer in a datacenter provides this information to the network/compute managers in the management plane. In some embodiments, the management plane collects contextual attributes from sources other than the host computer discovery and context engines. For instance, in some embodiments, the management plane collects from one or more servers compute context (e.g., cloud context from cloud vendors, or compute virtualization context by datacenter virtualization software), identity context from directory service servers, mobility context from mobility management servers, endpoint context from DNS (domain name server) and application inventory servers, network context (e.g., virtual network context from network virtualization server), etc.

By collecting the contextual information (e.g., information from the discovery and context engines and/or information from other context sources), the management plane can provide a user interface to the network/compute administrators to visualize the compute and network resources in the datacenter. Moreover, the collected contextual attributes allow the management plane to provide controls, through this user interface, for these administrators to specify context-based service rules and/or policies. These service rules/policies are then distributed to the host computers so that service engines on these computers can perform context-based service operations.

The preceding Summary is intended to serve as a brief introduction to some embodiments of the invention. It is not meant to be an introduction or overview of all inventive subject matter disclosed in this document. The Detailed Description that follows and the Drawings that are referred to in the Detailed Description will further describe the embodiments described in the Summary as well as other embodiments. Accordingly, to understand all the embodiments described by this document, a full review of the Summary, Detailed Description, the Drawings and the Claims is needed. Moreover, the claimed subject matters are not to be limited by the illustrative details in the Summary, Detailed Description and the Drawing.

BRIEF DESCRIPTION OF THE DRAWINGS

The novel features of the invention are set forth in the appended claims. However, for purposes of explanation, several embodiments of the invention are set forth in the following figures.

FIG. 1 illustrates a host computer that uses the context engine and context-based service engines of some embodiments of the invention.

FIG. 2 illustrates a more-detailed example of a host computer that in some embodiments is used to establish a distributed architecture for configuring and performing context-rich, attribute-based services in a datacenter.

FIG. 3 illustrates a process performed by a context engine of some embodiments.

FIG. 4 illustrates several examples of such firewall rules.

FIG. 5 illustrates several more detailed examples of the context-based firewall rules of some embodiments.

FIGS. 6-9 present various examples that illustrate the enforcement of various context-based firewall rule by a firewall engine.

FIG. 10 illustrates a process that the context engine performs to collect the user and group identifiers each time it receives a new network connection event from a GI agent.

FIG. 11 illustrates a process that a firewall engine performs in some embodiments.

FIG. 12 illustrates an example of how the service engines are managed in some embodiments.

FIG. 13 conceptually illustrates a computer system with which some embodiments of the invention are implemented.

DETAILED DESCRIPTION

In the following detailed description of the invention, numerous details, examples, and embodiments of the invention are set forth and described. However, it will be clear and apparent to one skilled in the art that the invention is not limited to the embodiments set forth and that the invention may be practiced without some of the specific details and examples discussed.

Some embodiments of the invention provide a novel architecture for capturing contextual attributes on host computers that execute one or more machines, and for consuming the captured contextual attributes to perform services on the host computers. Some embodiments execute a guest-introspection (GI) agent on each machine, from which contextual attributes need to be captured. In addition to executing one or more machines on each host computer, these embodiments also execute a context engine and one or more attribute-based service engines on each host computer. Through the GI agents of the machines on a host, the context engine of that host in some embodiments collects contextual attributes associated with network events and/or process events on the machines. The context engine then provides the contextual attributes to the service engines, which, in turn, use these contextual attributes to identify service rules that specify context-based services to perform on processes executing on the machines and/or data message flows sent by or received for the machines.

As used in this document, data messages refer to a collection of bits in a particular format sent across a network. One of ordinary skill in the art will recognize that the term data message may be used herein to refer to various formatted collections of bits that may be sent across a network, such as Ethernet frames, IP packets, TCP segments, UDP datagrams, etc. Also, as used in this document, references to L2, L3, L4, and L7 layers (or layer 2, layer 3, layer 4, layer 7) are references respectively to the second data link layer, the third network layer, the fourth transport layer, and the seventh application layer of the OSI (Open System Interconnection) layer model.

FIG. 1 illustrates a host computer 100 that uses the context engine and context-based service engines of some embodiments of the invention. As shown, host computer 100 includes several data compute nodes 105, a context engine 110, several context-based service engines 130, a threat detector 132, and a deep packet inspection (DPI) module 135. The context-based service engines include a discovery engine 120, a process control engine 122, an encryption engine 124, a load balancer 126 and a firewall engine 128. It also includes attribute-based, service-rule storages 140, and an attribute storage 145.

The DCNs are endpoint machines executing on the host computer 100. The DCNs are virtual machines (VMs) in some embodiments, containers in other embodiments, or a mix of VMs and containers in still other embodiments. On each DCN, a GI agent 150 executes in order to collect contextual attributes for the context engine 110. In some embodiments, the context engine 110 collects contextual attributes from the GI agents 150 of the DCNs on its host through a variety of different ways. For instance, in some embodiments, the GI agent on a DCN registers hooks (e.g., callbacks) with one or more modules (e.g., kernel-space modules or user-space modules) in the DCN's operating system for all the new network connection events and all the new process events.

Upon occurrence of a new network connection event, the GI agent 150 receives a callback from its DCN's OS and based on this callback, provides a network event identifier to the context engine 110. The network event identifier provides a set of attributes pertaining to the network event. These network event attributes in some embodiments include a five-tuple identifier (i.e., source port and IP address, destination port and IP address, and protocol) of the requested network connection, a process identifier of the process requesting the network connection, a user identifier associated with the requesting process, and a group identifier (e.g., an activity directory (AD) identifier) associated with the requesting process.

In some embodiments, the context engine directs the GI agent 150 to collect from the OS modules additional process parameters that are associated with the process identifier (ID) that it received with the network event. These additional process parameters in some embodiments include the process name, the process hash, the process path with command line parameters, the process network connection, the process-loaded modules, and one or more process consumption parameters specifying the process' consumption of one or more resources of the machine (e.g., central processing unit consumption, network consumption, and memory consumption). Instead of using the process identifier to query the GI agent 150 for additional process parameters associated with a network event, the context engine 110 in other embodiments receives all the process parameters associated with a network event in one shot, when the GI agent reports the network event to the context engine.

The OS of the DCN in some embodiments holds up a new network event (i.e., does not start sending data messages for the network event) until the GI agent 150 on that DCN directs it to proceed with processing the network event. In some of these embodiments, the GI agent 150 only allows the OS to proceed with processing the network event after the context engine 110 has collected all the needed attributes for this event (e.g., after receiving a message from the context engine that specifies that it has received all the process or network attributes that it needs for the new network event).

In some embodiments, the context engine 110 uses the process hash that it receives from the GI agent 150 to identify the name and version of the application (i.e., the software product) to which the process belongs. To do this, the context engine 110 in some embodiments, stores process hashes and associated application names/versions, compares the process hash that it receives from the GI agent with the stored process hashes to identify a matching hash, and then uses the application name/version of the matching hash as the application name and version of the process associated with the event.

In some embodiments, the context engine 110 obtains the process hashes and application names/versions from one or more network or compute managers, which may operate on another device or computer. In other embodiments, the context engine provides the hash associated with a process identifier to a network or compute manager, which then matches this hash to its process hash records and provides the application name/version of the associated process to the context engine. Once the context engine 110 obtains the application name/version associated with a network event, it can provide the name and version attributes to the attribute-based service engines, which can use this information (i.e., the application name and/or version) to identify the service rule to enforce.

Upon occurrence of a process event on a DCN 105, the DCN's GI agent 150 in some embodiments receives a callback from the DCN's OS and based on this callback, provides a process event identifier to the context engine 110. The process event identifier provides a set of attributes pertaining to the process event. This set of attributes includes the process identifier in some embodiments. In some embodiments, this set also includes a user identifier and a group identifier (e.g., an activity directory (AD) identifier).

In some embodiments, the GI agent provides all the process parameters (e.g., process identifier, user ID, group ID, process name, process hash, loaded module identifiers, consumption parameters, etc.) associated with a process event, to the context engine when it reports the process event to the context engine. In other embodiments, the context engine directs the GI agent to collect from the OS modules additional process parameters that are associated with the process identifier that context engine received with the process event. These additional process parameters in some embodiments are the same (e.g., process name, process hash, loaded module identifiers, consumption parameters, etc.) as the process parameters mentioned above for reported network events.

The context engine 110 of some embodiments augments the contextual attributes that it receives from the GI agents 150 with contextual attributes that it receives from other modules that execute on the host. The DPI module 135 (also referred to as the deep packet inspector) and the threat detector 132 (also referred to as the threat inspection module) are two such modules that provide contextual attributes to augment those that the context engine collects from the GI agents 150. In some embodiments, a DPI module is directed by the context engine 110 or another module (e.g., a firewall engine 128) to examine data messages of a data message flow associated with a process ID, to identify the type of traffic being sent in these data messages by the application associated with the process ID.

The identified traffic-type identity is today commonly referred to as the AppID. Also currently, there are a number of DPI modules that analyze messages of a data message flow to generate the AppID for the data message flow. In some embodiments, the context engine combines the AppID that it obtains for a network event with other context attributes that it identifies for this event, in order to produce a very rich set of attributes that the service engines can then use to perform their services. This rich set of attributes provides true application identity (i.e., application name, application version, application traffic type, etc.), based on which the service engines can perform their services. In some embodiments, the context engine 110 uses a network event's five-tuple identifier to associate the AppID for this event's data message flow with the contextual attributes that the context engine collects from the GI agent of the DCN associated with the data message flow (e.g., of the DCN from which the data message flow emanates).

The threat detector 132 provides a threat level indicator that specifies the threat level associated with a particular application that is executing on a DCN. Once the context engine obtains a set of process parameters that specify that a process has started on a machine or that is sending data messages on the machine, the context engine in some embodiments provides one or more process parameters (e.g., process hash, application name, application version, AppID, other process parameters, etc.) to the threat detection module.

This threat detection module then generates a threat level indicator (e.g., low, medium, high, etc.) for the identified process and provides this threat level indicator to the context engine. In some embodiments, the threat detector assigns a threat score to an application running on a DCN based on various application behavioral factors, such as (1) whether it does poor input validation, (2) whether it passes authentication credentials over unencrypted network links, (3) whether it uses weak password and account policies, (4) whether it stores configuration secrets in clear text, (5) whether it can transfer files, (6) whether the application is known to propagate malware, (7) whether the application is purposely evasive, (8) whether the application has known vulnerabilities, etc. In some embodiments, the threat detector is a third-party whitelisting application, such as the Bit9.

The context engine in some embodiments provides the threat level indicator, produced by the threat detector 132, to one or more service engines as another contextual attribute for performing services on a new process event or the data messages of a new network event; a service engine can use the threat level indicator as another attribute to identify service rules to enforce.

The context engine 110 stores the contextual attributes that it collects for network events and process events in the attribute storage 145. In some embodiments, the context engine stores each set of contextual attributes with one or more network event identifiers and/or process identifiers. For example, in some embodiments, the context engine 110 stores the collected contextual attributes for a new process event with the process identifier, or with a reference to this identifier. The context engine then uses the process identifier to provide the collected context attributes to a service engine (e.g., the process control engine 122) that performs a service for the process event.

The context engine in some embodiments stores the collected context attributes for a new network connection event with the five-tuple identifier of the network connection event, or with a reference to this five-tuple identifier. In some of these embodiments, the context engine provides to a service engine the context attributes for a network event along with this event's five-tuple identifier. The data messages for this network event will use this five-tuple identifier, and hence the service engine can use the supplied five-tuple identifier to identify the context attributes associated with a data message flow.

The context engine employs a push model in some embodiments to distribute the collected contextual attributes to the service engines 130, employs a pull model in other embodiments to distribute these attribute to the service engines 130, while in other embodiments, employing a push model for some service engines and a pull model for other service engines. In the push model, the context engine in some embodiments distributes to a service engine the contextual attributes that it collects for a process event or a network event with the process' identifier and/or the network event's flow identifier (e.g., the flow's five-tuple identifier).

In some embodiments, the context engine distributes to the service engine only the contextual attributes that are relevant for that service engine's service rules. In other words, in these embodiments, the context engine compares each collected attribute in a set of collected attributes (e.g., for a network event or a process event) with a list of attributes used by a service engine's service rules, and discards each collected attribute that is not used by the service rules. The context engine then provides to the service engine only the subset of collected attributes (in the set of collected attributes) that is being used by the engine's service rules. In other embodiments, the service engines perform this filtering operation to discard the contextual attributes that are not needed.

In the pull model, the context engine receives queries from a service engine for the contextual attributes that the context engine has collected for a particular process or network connection. In some embodiments, the context engine receives a process ID or a flow identifier (e.g., five-tuple identifier), with a query from the service engine, and uses the received identifier to identify the attribute set that it has to provide to the service engine.

In some embodiments, the context engine generates a service token (also called a service tag) for the collection of attributes that are relevant for the service engine, and provides this service token to another module (e.g., the GI agent or another module on the host) to pass along to the service engine (e.g., pass along in a data message's encapsulating tunnel header). The service engine then extracts the service token and provides this service token to the context engine, in order to identify the contextual attributes that the context engine has to provide to the service engine.

In some embodiments, the context engine 110 and the service engines 130 are all kernel space components of a hypervisor on which multiple VMs or containers execute, as further described below by reference to FIG. 2. In other embodiments, the context engine and/or one or more service engines are user space processes. For example, one or more service engines in some embodiments are service VMs (SVMs). In some embodiments, one or more service engines are in ingress datapaths and/or egress datapaths of DCNs, in order to receive access to data message flows to and from the DCNs to perform services on these data message flow. In other embodiments, one or more other modules on the host 100 intercept data messages from the ingress/egress datapaths and forward these messages to one or more service engines for these engines to perform services on the data messages. One such approach will be described below by reference to FIG. 2.

Different embodiments use different types of context-based service engines. In the example illustrated in FIG. 1, the service engines 130 include the discovery engine 120, the process control engine 122, the encryption engine 124, the load balancer 126, and the firewall engine 128. Each of these service engines 130 has an attribute-based, service-rule storage. FIG. 1 collectively represents all the context-based, service-rule storages of these service engines with the context-based, service rule storage 140 in order to simplify the illustration that is presented in this figure.

In some embodiments, each service rule in the service-rule storage 140 has a rule identifier for matching to a process or flow identifier to identify the rule to enforce for a process or network event. In some embodiments, the service rule storage 140 is defined in a hierarchical manner to ensure that a rule check will match a higher priority rule before matching a lower priority rule. Also, in some embodiments, the service rule storage 140 contains a default rule that specifies a default action for any rule check, as further explained below.

The firewall engine 128 performs firewall operations on data messages sent by or received for the DCNs 105. These firewall operations are based on firewall rules in the rule storage 140. Some of the firewall rules are defined purely in terms of layer 2-layer 4 attributes, e.g., in terms of five-tuple identifiers. Other firewall rules are defined in terms of contextual attributes that can include one or more of the collected contextual attributes, such as application names, application versions, AppID, resource consumption, threat level, user ID, group ID, etc. Yet other firewall rules in some embodiments are defined in terms of both L2-L4 parameters and contextual attributes. As the firewall engine 128 can resolve firewall rules that are defined by reference to contextual attributes, this firewall engine is referred to as a context-based firewall engine.

In some embodiments, the context-based firewall engine 128 can allow, block, or re-route data message flows based on any number of contextual attributes, because its firewall rules can be identified in terms of any combination of the collected contextual attributes. For example, this firewall engine can block all email traffic from chrome.exe when the user is part of a Nurse user group, and one firewall rule specifies that data messages should be blocked when the flow is associated with the Nurse group ID, the AppID identifies the traffic type as email, and the application name is Chrome. Similarly, context based firewall rules can block data message flows associated with video conferences, online video viewing, or use of old versions of software. Examples of such rules would block all Skype traffic, block all YouTube video traffic, block all HipChat audio/video conferences when application version number is older than a particular version number, block data message flows for any application with a high threat score, etc.

The load balancing engine 126 performs load balancing operations on data messages sent by the DCNs 105 to distribute the data message flows to different destination or service nodes in one or more destination/service node clusters. These load balancing operations are based on load-balancing rules in the rule storage 140. In some of these embodiments, each load-balancing rule can specify one or more load balancing criteria (e.g. a round robin criterion, a weighted round-robin criteria, etc.) for distributing traffic, and each criteria can be limited to a particular time range. In some embodiments, a load balancing operation involves replacing a data message flow's destination network address (e.g., the destination IP address, the destination MAC address, etc.) with another destination network address.

Some of the load-balancing rules are defined purely in terms of L2-L4 attributes, e.g., in terms of five-tuple identifiers. Other load-balancing rules are defined in terms of contextual attributes that can include one or more of the collected contextual attributes, such as application names, application versions, AppID, resource consumption, threat level, user ID, group ID, etc. Yet other load-balancing rules in some embodiments are defined in terms of both L2-L4 parameters and contextual attributes. As the load balancing engine 126 can resolve load-balancing rules that are defined by reference to contextual attributes, this load balancing engine is referred to as a context-based load balancer.

In some embodiments, the context-based load balancer 126 can distribute the data message flows based on any number of contextual attributes, because its load-balancing rules can be identified in terms of any combination of the collected contextual attributes. For example, the data distribution of the load balancer 126 can be based on any combination of user and application data. Examples of such load balancing operations include: (1) distributing data message flows associated with the Finance department on all load balancing pools, (2) redirecting all the Finance department's traffic to another pool when the primary pool for this department is down to make this department's traffic highly available, and (3) making all traffic associated with the Doctor's user group highly available. In some embodiments, the load-balancing rules can also be defined in terms of collected resource consumption, in order to distribute traffic to provide more or less resources to applications that consume a lot of resources on the DCNs.

The encryption engine 124 performs encryption/decryption operations (collectively referred to as encryption operations) on data messages sent by or received for the DCNs 105. These encryption operations are based on encryption rules in the rule storage 140. In some embodiments, each of these rules includes an encryption/decryption key identifier, which the encryption engine can use to retrieve an encryption/decryption key from a key manager on the host or operating outside of the host. Each encryption rule also specifies in some embodiments the type of encryption/decryption operation that the encryption module has to perform.

Each encryption rule also has a rule identifier. For some encryption rules, the rule identifiers are defined purely in terms of L2-L4 attributes, e.g., in terms of five-tuple identifiers. Other encryption rules are defined in terms of contextual attributes that can include one or more of the collected contextual attributes, such as application names, application versions, AppID, resource consumption, threat level, user ID, group ID, etc. Yet other encryption rules in some embodiments are defined in terms of both L2-L4 parameters and contextual attributes. As the encryption engine 124 can resolve encryption rules that are defined by reference to contextual attributes, this encryption engine is referred to as a context-based encryption engine.

In some embodiments, the context-based encryption module 124 can encrypt or decrypt the data message flows based on any number of contextual attributes, because its encryption rules can be identified in terms of any combination of the collected contextual attributes. For example, the encryption/decryption operation of the encryption engine 124 can be based on any combination of user and application data. Examples of such encryption operations include: (1) encrypt all traffic from Outlook (started on any machine) to Exchange Server, (2) encrypt all communication between applications in a three tier Webserver, Application Server and Database server, (3) encrypt all traffic originating from the Administrators Active Directory group, etc.

The process control engine 122 enforces context-based process control operations (e.g., process assessment and termination operations) on processes started on the DCNs 105. In some embodiments, whenever the context engine 110 receives a new process event from a GI agent 150 of a DCN, it provides the process parameters associated with this process event to the process control engine 122. This engine then uses the received set of process parameters to examine its service rule storage 140 to identify a matching context-based, process-control rule.

The process control engine 122 can direct the context engine to direct the GI agent of the DCN to perform a process-control operation on a process. Examples of such process-control operations include (1) terminating a video conference application that has a particular version number, (2) terminating a browser that is displaying YouTube traffic, (3) terminating applications that have a high threat level score.

The discovery engine 120 is another context-based service engine. In some embodiments, the discovery engine 120 captures new process events and new network events from the context engine, along with the contextual attributes that the context engine collects for these process and network events. As further described below, the discovery service engine then relays these events and their associated contextual attributes to one or more network managers (e.g., servers) that provide a management layer that allows network administrators to visualize events in a datacenter and specify policies for compute and network resources in the datacenter.

In relaying these events and attributes to the network management layer, the discovery module of some embodiments performs some pre-processing of these events and attributes. For example, in some embodiments, the discovery module filters some of the network or process events, while aggregating some or all of these events and their attributes. Also, in some embodiments, the discovery engine 120 directs the context engine 110 to collect additional contextual attributes for process or network events through the GI agents 150 or other modules (e.g., the DPI engine or threat detection engine), or to capture other types of events, such as file events and system events.

For example, in some embodiments, the discovery engine directs the context engine to build an inventory of the applications installed on the machines, and to periodically refresh this inventory. The discovery engine might so direct the context engine at the request of the management plane, or based on operational configurations that the management or control plane specifies for the discovery engine. In response to the request from the discovery engine, the context engine in some embodiments has each GI agent on each of its host's machine discover all installed processes on the machine, and all running processes and services.

After building an inventory of installed applications and the running processes/services, the discovery engine of a host computer in a datacenter provides this information to the network/compute managers in the management plane. In some embodiments, the management plane collects contextual attributes from sources other than the host computer discovery and context engines. For instance, in some embodiments, the management plane collects from one or more servers compute context (e.g., cloud context from cloud vendors, or compute virtualization context by datacenter virtualization software), identity context from directory service servers, mobility context from mobility management servers, endpoint context from DNS (domain name server) and application inventory servers, network context (e.g., virtual network context from network virtualization server), etc.

By collecting the contextual information (e.g., information from the discovery and context engines and/or information from other context sources), the management plane can provide a user interface to the network/compute administrators to visualize the compute and network resources in the datacenter. Moreover, the collected contextual attributes allow the management plane to provide controls through this user interface for these administrators to specify context-based service rules and/or policies. These service rules/policies are then distributed to the host computers so that service engines on these computers can perform context-based service operations.

In some embodiments described above, the same service engine 130 (e.g., the same firewall engine 128) performs the same type of service (e.g., a firewall service) based on service rules that can be defined in terms of message flow identifiers (e.g., five-tuple identifiers) or in terms of collected contextual attributes (e.g., AppID, threat level, user identifier, group identifier, application name/version, etc.) associated with the data message flows. In other embodiments, however, different service engines provide the same type of service based on the message flow identifiers (e.g., five-tuple identifiers) and based the collected contextual attributes of the data message flows. For instance, some embodiments use one flow-based firewall engine that performs firewall operations based on rules defined in terms of flow identifiers, and another context-based firewall engine that performs firewall operations based on rules defined in terms of context attributes (e.g., AppID, threat level, user identifier, group identifier, application name/version, etc.).

FIG. 2 illustrates a more-detailed example of a host computer 200 that in some embodiments is used to establish a distributed architecture for configuring and performing context-rich, attribute-based services in a datacenter. This host computer 200 includes many of the same components as host computer 100, such as context engine 110, service engines 130, threat detector 132, DPI module 135, attribute-based service rule storage 140, and context-attribute storage 145. Like in FIG. 1, the service engines 130 in FIG. 2 include the discovery engine 120, the process control engine 122, the encryption engine 124, the load balancer 126 and the firewall engine 128.

In FIG. 2, the DCNs are VMs 205 that execute on a hypervisor. Also, in FIG. 2, the host computer 200 includes a software forwarding element 210, an attribute-mapping storage 223, a connection state data storages 225, a multiplexer (MUX) 227, and a context-engine policy storage 143. In some embodiments, the context engine 110, the software forwarding element 210, the service engines 130, the rule storages 140, the connection state data storage 225, the context-engine policy storage 143, and the MUX 227 operate in the kernel space of the hypervisor, while the VMs 205 operate in the hypervisor's user space. In other embodiments, one or more service engines are user space modules (e.g., are service VMs).

In some embodiments, the VMs 205 serve as data end points in the datacenter. Examples of such machines include webservers, application servers, database servers, etc. In some cases, all the VMs belong to one entity, e.g., an enterprise that operates the host. In other cases, the host 200 operates in a multi-tenant environment (e.g., in a multi-tenant data center), and different VMs 205 may belong to one tenant or to multiple tenants.

Each VM 205 includes a GI agent 250 that interacts with the context engine 110 to provide context attribute sets to this engine, and to receive instructions and queries from this engine. These interaction between the GI agents 250 and the context engine 110 are similar to the interactions described above between the GI agents 150 and the context engine 110. However, as shown in FIG. 2, all the communication between the context engine 110 and the GI agents 250 in some embodiments are relayed through the MUX 227. One example of such a mux is the mux that is used by the Endpoint Security (EPSec) platform of ESX hypervisors of VMware Inc.

In some embodiments, the GI agents communicate with the MUX 227 through a fast communication channel (such as VMCI channel of ESX). In some embodiments, this communication channel is a shared memory channel. As mentioned above, the attributes collected by the context engine 110 from the GI agents 250 in some embodiments include a rich group of parameters (e.g., layer 7 parameters, process identifiers, user identifiers, group identifiers, process name, process hash, loaded module identifiers, consumption parameters, etc.)

As shown, each VM 205 also includes a virtual network interface card (VNIC) 255 in some embodiments. Each VNIC is responsible for exchanging messages between its VM and the software forwarding element (SFE) 210. Each VNIC connects to a particular port 260 of the SFE 210. The SFE 210 also connects to a physical network interface card (NIC) (not shown) of the host. In some embodiments, the VNICs are software abstractions created by the hypervisor of one or more physical NICs (PNICs) of the host.

In some embodiments, the SFE 210 maintains a single port 260 for each VNIC of each VM. The SFE 210 connects to the host PNIC (through a NIC driver (not shown)) to send outgoing messages and to receive incoming messages. In some embodiments, the SFE 210 is defined to include a port 265 that connects to the PNIC's driver to send and receive messages to and from the PNIC. The SFE 210 performs message-processing operations to forward messages that it receives on one of its ports to another one of its ports. For example, in some embodiments, the SFE tries to use data in the message (e.g., data in the message header) to match a message to flow based rules, and upon finding a match, to perform the action specified by the matching rule (e.g., to hand the message to one of its ports 260 or 265, which directs the message to be supplied to a destination VM or to the PNIC).

In some embodiments, the SFE 210 is a software switch, while in other embodiments it is a software router or a combined software switch/router. The SFE 210 in some embodiments implements one or more logical forwarding elements (e.g., logical switches or logical routers) with SFE executing on other hosts in a multi-host environment. A logical forwarding element in some embodiments can span multiple hosts to connect VMs that execute on different hosts but belong to one logical network.

Different logical forwarding elements can be defined to specify different logical networks for different users, and each logical forwarding element can be defined by multiple software forwarding elements on multiple hosts. Each logical forwarding element isolates the traffic of the VMs of one logical network from the VMs of another logical network that is serviced by another logical forwarding element. A logical forwarding element can connect VMs executing on the same host and/or different hosts. In some embodiments, the SFE extracts from a data message a logical network identifier (e.g., a VNI) and a MAC address. The SFE in these embodiments uses the extracted VNI to identify a logical port group, and then uses the MAC address to identify a port within the port group.

Software switches (e.g., software switches of hypervisors) are sometimes referred to as virtual switches because they operate in software and they provide the VMs with shared access to the PNIC(s) of the host. However, in this document, software switches are referred to as physical switches because they are items in the physical world. This terminology also differentiates software switches from logical switches, which are abstractions of the types of connections that are provided by the software switches. There are various mechanisms for creating logical switches from software switches. VXLAN provides one manner for creating such logical switches. The VXLAN standard is described in Mahalingam, Mallik; Dutt, Dinesh G.; et al. (2013 May 8), VXLAN: A Framework for Overlaying Virtualized Layer 2 Networks over Layer 3 Networks, IETF.

The ports of the SFE 210 in some embodiments include one or more function calls to one or more modules that implement special input/output (I/O) operations on incoming and outgoing messages that are received at the ports. Examples of I/O operations that are implemented by the ports 260 include ARP broadcast suppression operations and DHCP broadcast suppression operations, as described in U.S. Pat. No. 9,548,965. Other I/O operations (such as firewall operations, load balancing operations, network address translation operations, etc.) can be so implemented in some embodiments of the invention. By implementing a stack of such function calls, the ports can implement a chain of I/O operations on incoming and/or outgoing messages in some embodiments. Also, in some embodiments, other modules in the data path (such as the VNICs 255, port 265, etc.) implement the I/O function call operations, instead of, or in conjunction with, the ports 260.

In some embodiments, one or more of function calls of the SFE ports 260 can be to one or more service engines 130 that process context-based service rules in the context based service rule storages 140. Each service engine 130 in some embodiments has its own context-based service rule storage 140, attribute mapping storage 223, and connection cache storage 225. FIG. 2 presents just one service rule storage 140, attribute mapping storage 223, and connection cache storage 225 for all the service engines, in order not to obscure the presentation in this figure with unnecessary detail. Also, in some embodiments, each VM has its own instance of each service engine 130 (e.g., its own instance of discovery engine 120, process control engine 122, encryption engine 124, load balancer 126, and firewall engine 128). In other embodiments, one service engine can service data message flows for multiple VMs on a host (e.g., VMs for the same logical network).

To perform its service operation for a data message flow, a service engine 130 in some embodiments tries to match the flow identifier (e.g., the five-tuple identifier) and/or the flow's associated context attribute set to the rule identifiers of its service rules in its service rule storage 140. Specifically, for a service engine 130 to perform its service check operation for a data message flow, the SFE port 260 that calls the service engine supplies a set of attributes of a message that the port receives. In some embodiments, the set of attributes are message identifiers, such as traditional five-tuple identifiers. In some embodiments, one or more of the identifier values can be logical values that are defined for a logical network (e.g., can be IP addresses defined in a logical address space). In other embodiments, all of the identifier values are defined in the physical domains. In still other embodiments, some of the identifier values are defined in the logical domain, while other identifier values are defined in the physical domain.

The service engine in some embodiments then uses the received message's attribute set (e.g., the message's five-tuple identifier) to identify the context attribute set that the service engine has stored for this flow in the attribute-mapping storage 223. As mentioned above, the context engine 110 in some embodiments supplies the context attributes for new flows (i.e., new network connection events) and for new processes to the service engines 130, along with a flow identifier (e.g., a five-tuple identifier) or a process identifier. The context-engine policy storage 143 contains the rules that control the operation of the context engine 110. In some embodiments, these policies direct the context engine to generate rules for the service engines or to direct the service engines to generate rules (e.g., when a high-threat application runs on a VM, directing the encryption engine for all the other VMs on the same host to encrypt their data message traffic). The service engines 130 in these embodiments store the context attributes that they receive from the context engine in the attribute-mapping storage 223.

In some embodiments, a service engine 130 stores the context attribute set for each new flow or new process with that flow's identifier (e.g., five-tuple identifier) or that process' identifier in the attribute-mapping storage. In this manner, the service engine can identify the context attribute set for each new flow that it receives from the SFE port 260 by searching its attribute-mapping storage 223 for a context record that has a matching flow identifier. The context record with the matching flow identifier includes the context attribute set for this flow. Similarly, to identify the context attribute set for a process event, a service engine in some embodiments searches its attribute-mapping storage 223 for a context record with a matching process identifier.

As mentioned above, some or all of the service engines in some embodiments pull the context attribute sets for a new flow or new process from the context engine. For instance, in some embodiments, a service engine supplies a new flow's five-tuple identifier that it receives from the SFE port 260, to the context engine 110. This engine 110 then examines its attribute storage 145 to identify a set of attributes that is stored for this five-tuple identifier, and then supplies this attribute set (or a subset of it that it obtains by filtering the identified attribute set for the service engine) to the service engine.

As described above, some embodiments implement the pull model by using a service token to encode the attribute set for a new message flow. When notified of a new network connection event, the context engine 110 in some embodiments (1) collects the context attribute set for the new event, (2) filters this set to discard the attributes that are not relevant for performing one or more services on the flow, (3) stores the remaining filtering attribute subset in the attribute storage 145 along with a service token, (4) provides the service token to the GI agent 250. The GI agent then passes or has another module pass this token to be passed to the service engine(s) in-band (e.g., in a tunnel header of a data message that the agent's VM sends to a destination) or out-of-band (i.e., separately from the data messages that the agent's VM sends to a destination).

When the service engine gets the new flow through the SFE port 260, it supplies this flow's service token to the context engine, which uses this service token to identify in its context-attribute storage 145 the context attributes to supply to the service engine. In the embodiments that the SFE port does not provide this service token to the service engine, the service engine first has to identify the service token by searching its data stores using the flow's identifier before supplying the service token to the context engine.

After identifying the contextual attribute set for a data message flow, the service engine 130 in some embodiments performs its service operation based on service rules that are stored in the service rule storage 140. To perform its service operation, the service engine 130 matches the received attribute subset with corresponding attribute sets that are stored for the service rules. In some embodiments, each service rule in the rule storage 140 has a rule identifier and an action parameter set.

As mentioned above, the rule identifier of a service rule in some embodiments can be defined in terms of one or more contextual attributes that are not L2-L4 header parameters (e.g., are L7 parameters, process identifiers, user identifiers, group identifiers, process name, process hash, loaded module identifiers, consumption parameters, etc.). In some embodiments, a rule identifier can also include L2-L4 header parameters. Also, in some embodiments, one or more parameters in a rule identifier can be specified in terms of an individual value or a wildcard value. Also, in some embodiments, a rule identifier can include a set of individual values or a group identifier, such as a security group identifier, a compute construct identifier, a network construct identifier, etc.

To match a received attribute set with the rules, the service engine compares the received attribute set with the associated identifiers of the service rules stored in the service rule storage 140. Upon identifying a matching rule, the service engine 130 performs a service operation (e.g., a firewall operation, a load balancing operation, an encryption operation, other middlebox operation, etc.), based on the action parameter set (e.g., based on Allow/Drop parameters, the load balancing criteria, encryption parameters, etc.) of the matching rule.

In some embodiments, the service rule storage 140 is defined in a hierarchical manner to ensure that when the message's attribute subset matches multiple rules, a message rule check will match a higher priority rule before matching a lower priority rule. Also, in some embodiments, the service rule storage 140 contains a default rule that specifies a default action for any message rule check that cannot identify any other service rules; this default rule will be a match for all possible attribute subsets in some embodiments, and ensures that the service rule engine will return an action for all received attribute subsets. In some embodiments, the default rule will specify no service.

Multiple messages can have the same message identifier attribute sets, e.g., when the messages are part of one flow that is associated with one communication session between two machines. Accordingly, after matching a data message with a service rule in the storage 140 based on the message's identified context attribute set, the service engine of some embodiments stores the service rule (or a reference to the service rule) in the connection state cache storage 225, so that it can later use this service rule for subsequent data messages of the same flow.

In some embodiments, the connection state cache storage 225 stores the service rule, or a reference to the service rule, that the service engine 130 identifies for different message identifier sets (e.g., for different five-tuple identifiers that identify different data message flows). In some embodiments, the connection state cache storage 225 stores each service rule, or reference to the service rule, with an identifier (e.g., a flow's five-tuple identifier and/or a hash value of the flow's five-tuple identifier) that is generated from the matching message identifier set.

Before checking with the service rule storage 140 for a particular message, the service rule engine 130 of some embodiments checks the connection state data storage 225 to determine whether this storage has previously identified a service rule for this message's flow. If not, the service engine 130 identifies the contextual attribute set for the message flow, and then checks the service rule storage 140 for a service rule that matches the message's identified attribute set and/or its five-tuple identifier. When the connection state data storage has an entry for the particular message, the service engine performs its service operation based on this service rule's action parameter set.

In the service architecture of FIG. 2, the DPI module 135 performs deep packet inspection on a data message flow at the direction of the firewall engine 128. Specifically, when the firewall engine 128 receives a new data message that is part of a new data message flow, the firewall engine in some embodiments directs the DPI module to inspect that new data message and one or more of the next few data messages in the same flow. Based on this examination, the DPI engine identifies the type of traffic (i.e., the application on the wire) that is being sent in this data message flow, generates an AppID for this traffic type, and stores this AppID in the context-attribute storage 145. In some embodiments, the context attribute sets are stored in the attribute storage based on flow identifiers and/or process identifier. Accordingly, in some embodiments, the DPI engine 135 stores the AppID for a new data message flow in the context-attribute storage 145 based on that flow's five-tuple identifier.

In some embodiments, the context engine 110 pushes to the service engines 130 the AppID for a new data message flow once the DPI engine stores the AppID in the attribute storage 145. In other embodiments, the context engine 110 pulls the AppID from the context-attribute storage 145 whenever it is queried for the contextual attributes for a data message flow by a service engine. In some embodiments, the context engine 110 uses the five-tuple identifier of the flow to identify the record in the attribute storage 145 with the matching record identifier and the AppID.

FIG. 3 illustrates a process 300 that the context engine 110 performs in some embodiments each time it is notified about a new process or network-connection event. From a GI agent 250 of a VM 205, the process 300 initially receives (at 305) a notification regarding a new process or network connection event. Next, at 310, the process 300 collects all the desired contextual attributes regarding the notified event.

As described above, the context engine 110 in some embodiments interacts (at 310) with the reporting GI agent 250 to collect additional information regarding a reported event. The GI agent in some embodiments interacts with the network stack and/or process subsystem in the VM's OS kernel space to collect contextual attributes regarding a process or network event. The GI agent in some embodiments also collects this information from user-space modules (e.g., a user mode dynamic link library, DLL) that operate in user-space process (e.g., a VMtool.exe) to collect contextual attributes. On VM's using Microsoft Windows, the GI agent in some embodiments registers hooks in the Windows Filtering Platform (WFP) to obtain network events, while registering in the Window's Process Subsystem to collect process related attributes. In some embodiments, the GI agent hook is at the Application Layer Enforcement (ALE) layer of WFP, so that it can capture all socket-connection requests from application processes on the VM.

In some embodiments, the context engine 110 interacts with the management or control plane to collect contextual attributes, and/or to receive records that it can examine to identify contextual attributes for identified network or process events. In some of these embodiments, the context engine interacts with a management or control plane proxy (that operates on its host) in order to obtain data from the management or control plane servers that operate outside of the host. In some of these embodiments, the context engine operates in the kernel space.

After collecting the contextual attributes at 310, the process uses (at 315) the attributes of the received event or the contextual attributes collected for the received event to identify one or more policies in the context-engine policy storage 143. At 315, the process identifies any policy that has a policy identifier that matches the collected attributes and event attributes.

Next, at 320, the process produces context-attribute mapping records for one or more service engines based on the policies identified at 315. One or more of the identified policies might specify that for a particular process or network event a particular set of service engines need to be notified about the event (e.g., about a new data message flow), with each service engine receiving a subset of contextual attributes that are relevant for that service engine, to perform its processing for that event. This operation in some embodiments involves the context engine not including attributes that are not relevant for each particular service engine in the subset of contextual attributes that it provides to that particular service engine.

In some embodiments, certain events might necessitate new service rules to be created for one or more service engines. For example, when a high-threat application is identified on one VM, a policy might specify that other VMs on that host might have to start to encrypt their data message traffic. In some such embodiments, the policy storage 143 includes policies that direct the context engine to generate service rules for service engines under certain circumstances, or to direct the service engines to generate such service rules. For such embodiments, the process (at 320) if needed generates service rules for service engines under certain circumstances, or directs the service engines to generate such service rules

At 325, the process 300 distributes the mapping records and/or generated service rules/instructions to one or more service engines. As mentioned above, the context engine can employ a push model or a pull model to distribute such records and/or rules/instructions. When employing a pull model, the process 300 in some embodiments not only performs the operation 325 in response to a query from a service engine, but also performs some or all of the operation 320 in response to this query. After 325, the process ends.

The firewall engine 128 is a context-based firewall engine that performs its firewall operations based on firewall rules that can be specified in terms of not only L2-L4 parameters, but also in terms of contextual attributes. FIG. 4 illustrates several examples of such firewall rules. This figure illustrates a firewall rule storage 140 of some embodiments. As shown, each firewall rule includes a rule identifier 405 and a firewall action 410.

In some embodiments, a firewall action parameter 410 can be any one of the traditional firewall actions, such as Allow, Drop, Re-Route, etc. Each rule identifier 405 specifies one or more data tuples that can be used to identify a rule that matches a data message flow. As shown, a rule identifier can include in some embodiments any L2-L4 parameters (e.g., source IP address, source port, destination port, protocol, etc.). One or more of these parameters can be virtual parameters (e.g., a VIP of a destination cluster) or a logical identifier (e.g., a logical network identifier).

In some embodiments, a rule identifier can also include contextual attributes, such as AppID, application name, application version, user ID, group ID, threat level, and resource consumption. In some embodiments, a firewall engine searches a firewall data storage by comparing one or more message attributes (e.g., five-tuple header values, contextual attributes) with the rule identifiers 405 to identify the highest priority rule with a matching rule identifier.

In some embodiments, different firewall engines 128 on different hosts enforce the same set of firewall rules. For instance, in some embodiments, different firewall engines 128 process the same firewall rules on different hosts for VMs of one logical network in order to provide a level of security on data messages that are transmitted by or received for these VMs. For this logical network, these firewall engines 128 collectively form a distributed firewall engine (i.e., a single, conceptual logical firewall engine) that spans across the multiple hosts.

FIG. 5 illustrates several more detailed examples of the context-based firewall rules of some embodiments. In these examples, the rule identifier 405 of each rule is expressed in terms of the five-tuple identifier and one or more contextual attributes. Each rule has one or more attributes in its five-tuple identifier that are wildcard values designated by an asterisk in order to specify that the value of these attributes do not matter (i.e., the data message flow can have any value for these attributes without failing to match the rule).

The first rule 535 specifies that all data message flows from Skype version 1024 should be dropped. The rule identifier for this rule is expressed only in terms of the contextual attributes of the data message flow. As mentioned above, and further described below, each time the firewall engine 128 identifies a new data message flow, it identifies the flow's contextual attributes by interacting with the context engine or by examining the records in its attribute-mapping storage 223 to identify a record that specifies the contextual attributes for the flow's five-tuple identifier.

The second rule 530 specifies that all data message flows that have a Group ID equal to Nurses and AppID equal to YouTube traffic are to be dropped. By enforcing this rule, the firewall engine 128 can make sure that nurses that login to its VM 205 cannot view YouTube traffic. Again, the rule identifier for this rule is expressed only in terms of the contextual attributes of the data message flow. In this example, the contextual attributes are the Group ID and the AppID.

FIG. 6 presents an example that illustrates the enforcement of the second rule 530 by the firewall engine 128. Specifically, it shows this firewall engine 128 allowing a first data message 605 from a browser 610 to pass through, while blocking a second data message 615 from this browser. As shown, both these data messages are associated with an operation that a nurse 620 has performed on the browser. The first data message flow is allowed to pass through as it relates to an email that the nurse is sending through the browser. This firewall engine 128 allows this message to go through because it does not match any firewall rule that requires the message to be blocked. The firewall engine 128, on the other hand, blocks the second data message as it relates to the nurse trying to watch a YouTube video, and this type of data message flow is prohibited by rule 530.

The third rule 525 in FIG. 5 specifies that all data message flows that are associated with a high threat level indicator should be blocked if they are going to a particular destination IP address A. The rule identifier for this rule is defined in terms of a contextual attribute (i.e., the high threat level indicator) and one attribute (the destination IP address) in the five-tuple identifier of the data message.

FIG. 7 presents an example that in two stages illustrates the enforcement of this rule 525 by the firewall engine 128. Specifically, it shows in a first stage 702 the firewall engine 128 allows data messages 710 to pass from a VM 205 to another VM 720 (outside of the host) that has the particular destination IP address A. In the second stage 704, an application 705 is installed on the VM 205. This application is designated as a high threat application by the threat detector 132. Whenever a new data message flow starts on the VM, the context engine associates this data message flow with a high threat level tag. Accordingly, in the second stage 704, the firewall engine 128 blocks a data message 750 from the VM 205 to the other VM 720 as this data message is associated with the high threat level, and the rule 525 prohibits such a data message to be sent to IP address A of the VM 720.

The fourth and fifth rules 520 and 515 in FIG. 5 specify that data messages associated with the Doctor and Nurses groups can access VMs associated with a VIP address A, while data messages associated with the Accountant group cannot access these VMs. The rule identifier for the fourth rule is defined in terms of two contextual attributes (i.e., Doctor and Nurse group identifiers) and one attribute (the VIP destination address A) in the five-tuple identifier of the data message. The rule identifier for the fifth rule is defined in terms of one contextual attribute (i.e., Accountant group identifier) and one attribute (i.e., the VIP destination address A) in the five-tuple identifier of the data message. In some embodiments, the VIP address is an address of a cluster of VMs that perform the same function, and a load balancer will translate this VIP address to the IP address of one of the VMs in the cluster.

FIG. 8 presents an example that illustrates the enforcement of the fourth and fifth rule 520 and 515 by the firewall engine 128. Specifically, it shows two users concurrently logged into one VM that is acting as a terminal server. One of these users is a Nurse X, while another user is an Accountant Y. FIG. 8 further shows that the firewall engine 128 allows a first data message 805 from the Nurse X's session to pass through to a VM in a VM cluster 850 identified by the destination IP address VIP A. It also shows the firewall engine blocking a second data message 810 from the Accountant Y's session from reaching any of the VMs in this VM cluster, because the fifth rule 515 prevents data messages that are associated with the Accountants Group ID to reach the destination IP address VIP A.

In FIG. 8, the two data messages are for two different actual users who are concurrently logged into a VM. In other cases, only one user might be actually logged in to a VM, but administrative processes might be running on the VM, in conjunction with the processes that run for the applications started by the logged in user. The administrative processes can be services/daemons running in a VM in a different user context than the logged in user. Services generally run in admin/root context and not in the logged in user context. This is a potential security hole as it might allow any application running in a non-logged on user context to access a network resource. Accordingly, even when only a single user is logged in to a VM, it can be desirable to specify firewall rules that treat differently the data messages associated with background administrative processes from data messages associated with processes that run for the applications started by the logged in user.

As an example, the sixth rule 510 allows data messages of processes associated with the applications operated by an individual in the High Security Group to access other VMs with sensitive data (in this example, these VMs are part of a cluster of VMs associated with an IP address VIP B), while the seventh rule 505 blocks data messages associated with background administrative processes from accessing such VMs. This is useful for ensuring that IT personnel or hackers cannot create a back door access to sensitive data by installing administrative processes that access high security VMs that piggyback off the login session of a user with the appropriate clearance.

FIG. 9 presents an example that illustrates the enforcement of the sixth and seventh rules 510 and 505 by the firewall engine 128. Specifically, it shows two data message flows concurrently emanating from one VM. One data message flow is associated with the CEO, while another is associated with a background IT utility process called Utility Q. FIG. 9 further shows that the firewall engine 128 allows a data message 905 from the CEO's session to pass through to a VM in a high security cluster 950 identified by the VIP address B. It also shows the firewall engine blocking a second data message 910 from the Utility Q′s session from reaching any of the VMs in the high security cluster, because the seventh rule 505 prevents data messages that are associated with administrative processes (such as the Utility Q process) from reaching the VIP address B.

The firewall engine can differentiate the data message flows for two different processes that run concurrently on a VM for two different login/administrative credentials because the context engine collects the user and group identifiers for each data message flow when each flow starts, and associates each flow with its user and group identifiers. FIG. 10 illustrates a process 1000 that the context engine performs to collect the user and group identifiers each time it receives a new network connection event from a GI agent.

The process 1000 initially receives (at 1005) a notification regarding a new network connection event from a GI agent 250 on a VM 205. As mentioned above, the GI agent in some embodiments provides the following information in a new network connection notification: the connection's five-tuple identifier, the identifier of the process requesting the network connection, a user identifier associated with the requesting process, and a group identifier associated with the requesting process.

Next, at 1010, the process 1000 queries the GI agent to collect any other contextual attributes needed for the new network connection event. Examples of such additional parameters include additional parameters associated with the process requesting the network connection. At 1015, the process 1000 then publishes one or more context attribute records to one or more service engines. In some embodiments, each context attribute record to each service engine includes the connection's five-tuple identifier and a set of one or more context attributes, including the user identifier and/or the group identifier. The service engines then store the provided contextual attribute records in their mapping storages 223, so that they can use these records to identify the context attribute sets associated with different data message flows that they process. For service engines that have different service rules for different processes concurrently running on a VM for different user accounts, the context attribute sets include a user identifier or a group identifier to allow these service engines to associate different data message flows from the VM to different user/group identifiers.

In some embodiments, the context engine does not include in a service engine's context attribute record the contextual attributes that are not needed by the service engine. Also, in some embodiments, the context engine provides different context attribute records to different context engines for the same network connection event, because different service engines need different sets of context attributes. As described above, the context engine 110 in some embodiments does not push the context attribute sets for new network connections to some or all of the service engines, but rather has these service engines pull these attribute sets.

In some embodiments, the context engine can associate a data message flow on a source host with the source VM (i.e., the VM that is the source of the data message flow) with a contextual attribute of a destination VM on the same host or a different destination host. The firewall engine 128 in these embodiments can then use such destination-based contextual attributes to resolve firewall rules. For instance, the firewall engine can drop all data messages addressed to a particular type of server (e.g., a Sharepoint server). To support such destination-based rules, the context engines of some embodiments directs the GI agents to identify the processes that register for notifications on particular ports, and uses this information along with process identifiers and hashes to identify the applications that serve as destinations of data message flows. The information collected by the context engines on the different hosts are collected by the management plane (e.g., by the management servers that operate on separate computers or on the same hosts that execute the VMs), which aggregates this data and distributes the aggregated data to the other context engines. The distributed information can then be used by the context engines on the host to resolve contextual policies on these hosts, in order to supply context-based rules to the context-based service engines on these hosts.

FIG. 11 illustrates a process 1100 that the firewall engine 128 performs in some embodiments. As shown, the process 1100 starts when the firewall engine receives (at 1105) a data message from its corresponding SFE port 260. This port relays this message when it receives the data message from its VM or for its VM. In some embodiments, the port relays the data message by passing to the firewall engine a reference (e.g., a handle that identifies a location in memory that stores the data message) to the data message or the data message's header values.

The process determines (at 1110) whether the connection state cache 225 stores a record that identifies a firewall action for the message flow of the received data message. As mentioned above, each time a firewall engine uses a firewall rule to process a new data message, the firewall engine in some embodiments creates a record in the connection state cache 225 to store the firewall action performed, so that when the firewall engine receives another data message within the same flow (i.e., with the same five-tuple identifier), it can perform the same firewall action that it performed on previous data messages in the same flow. The use of the connection state cache 225 allows the firewall engine 128 to process the data message flows more quickly. In some embodiments, each cached record in the connection state cache 225 has a record identifier that is defined in terms of data message identifiers (e.g., five-tuple identifiers). In these embodiments, the process compares the received data message's identifier (e.g., five-tuple identifier) with the record identifiers of the cached records to identify any record with a record identifier that matches the received data message's identifier.

When the process 1100 identifies (at 1110) a record for the received data message's flow in the cache 225, the process (at 1115) then performs the firewall action (e.g., Allow, Drop, Re-Route, etc.) specified in this record. Assuming that the firewall action does not require the data message to be dropped, the process 1100 sends the processed data message along its datapath. In some embodiments, this operation entails returning a communication to the SFE port 260 (that called the firewall engine to initiate the process 1100) to indicate that the firewall engine is done with its processing of the VM data message. The SFE port 260 can then handoff the data message to the SFE or the VM, or can call another service engine in the I/O chain operator to perform another service operation on the data message.

When the firewall action performed at 1115 results in the dropping of the data message, the process 1100 notifies (at 1115) the SFE port 260 of this operation. Also, when the firewall action performed at 1115 requires the data message to be re-routed, the process 1100 performs (at 1115) a network address translation on the data message in order to effectuate this re-routing, and then returns (at 1115) the data message to the SFE port so that the data message can be sent along its datapath. After 1115, the process 1100 ends.

When the process 1100 determines (at 1110) that the connection cache 225 does not store a record for the received data message's flow, the process 1100 identifies (at 1120) one or more contextual attributes for this data message flow. As mentioned above, the service engines of different embodiments perform this operation differently. For instance, in some embodiments, the firewall engine 128 checks the mapping storage 223 for a record that has a record identifier that matches the received data message's header values (e.g., its five-tuple identifier). It then uses (at 1120) the contextual attribute set of this matching record as the contextual attribute set of the received data message flow.

In other embodiments, the firewall engine 128 queries the context engine to obtain the contextual attribute set for the received data message. With this query, the firewall engine supplies the received message's flow identifier (e.g., five-tuple identifier) or its associated service token. The context engine then uses the message's flow identifier or its associated service token to identify a set of contextual attributes in its context attribute storage 145, as explained above.

Once the process 1100 has obtained the contextual attribute set for the received data message, it uses this attribute set along with the message's other identifiers to identify (at 1125) a firewall rule in the firewall rule data store 140 for the data message received at 1105. For instance, in some embodiments, the firewall rules have rule identifiers that are defined in terms of one or more of the five-tuple attributes along with one or more contextual attributes, such as application name, application version, user ID, group ID, AppID, threat level, resource consumption level, etc. To identify the firewall rule in the data store 140, the process in some embodiments compares the contextual attributes and/or other attributes (e.g., five-tuple identifier) of the received data message with the rule identifiers (e.g., rule identifiers 405) of the firewall rules to identify the highest priority rule that has an identifier that matches the message's attribute set.

In some embodiments, the process uses different message-attribute sets to perform this comparison operation. For instance, in some embodiments, the message attribute set includes the destination IP address of the message (e.g., the VIP of the addressed node group) along with one or more contextual attributes. In other embodiments, the message attribute set includes other attributes, such as one or more of the other five-tuple identifiers (e.g., one or more of the source IP, source port, destination port, and protocol). In some embodiments, the message attribute set includes logical network identifiers such as virtual network identifier (VNI), virtual distributed router identifier (VDRI), a logical MAC address, a logical IP address, etc.

When the process identifies a firewall rule (at 1125), it performs the firewall action (e.g., Allow, Drop, Re-Route, etc.) of this rule on the received data message. Assuming that the firewall action does not require the data message to be dropped, the process 1100 sends (at 1130) the processed data message along its datapath. In some embodiments, this operation entails returning a communication to the SFE port 260 (that called the firewall engine to initiate the process 1100) to indicate that the firewall engine is done with its processing of the VM data message. The SFE port 260 can then handoff the data message to the SFE or the VM, or can call another service engine in the I/O chain operator to perform another service operation on the data message.

When the firewall action performed at 1130 results in the dropping of the data message, the process 1100 notifies (at 1130) the SFE port 260 of this operation. Also, when the firewall action performed at 1130 requires the data message to be re-routed, the process 1100 performs (at 1130) a network address translation on the data message in order to effectuate this re-routing, and then returns (at 1130) the data message to the SFE port so that the data message can be sent along its datapath. After performing the firewall action at 1130, the process creates (at 1135) a record in the connection cache data store 225. This record identifies the firewall action for the received data message's flow. After 1135, the process ends.

In some embodiments, the management plane obtains an inventory of all processes and services that are running on the VMs on the hosts in a datacenter. The discovery engine 120 of a host 200 in some embodiments assists in collecting this data from the VMs executing on its host. In some embodiments, the inventoried process/services are referred to as the inventoried applications, which include all client processes, services or daemons that utilize network input/output and all server processes that have registered to listen to (i.e., to obtain messages) certain network connections. The discovery engine collects this data using the GI agents 250 and the MUX 227 in some embodiments.

Based on the data collected by all the discovery engines on all the hosts, the management servers (e.g., the network managers and/or compute managers) build the inventory of the running applications. In some embodiments, each application is identified by comparing its file hash obtained from the VMs 205 with hashes of application files stored in the application data storage of the management plane. The management plane in some embodiments has the discovery engines update their data collection so that the management plane can refresh its inventor on a scheduled basis.

The management plane in some embodiments then provides a rule creation interface for allowing administrators to create context-based firewall rules and/or policies for the firewall engines 128 (as well as service rules for the other service engines 130). The rule creation interface allows the administrators to define high-level firewall policies (and other service policies) based on applications inventoried through the data collected by the discovery engines 120, and contextual attributes collected by the context engines 110 and by the management plane's interface with other management server clusters.

Once the high-level firewall policies (and other service policies) are defined in the management plane, the management plane directly supplies some or all of these policies to the management proxies (not shown) on the hosts 200, and/or indirectly supplies some or all of these policies to these proxies through a set of controllers (e.g., network controllers). In some embodiments, the management proxies publish the received policies as rules to the service engine rule storages 140. In some embodiments, the proxies transform these policies before publishing them to the service rule storages 140. For instance, in some embodiments, the policies are published with AppliedTo tuples that identify the service nodes and/or logical networks to which they are associated. In some of these embodiments, the management proxies on the hosts remove the AppliedTo tuple from each service policy, before pushing the policy as a service rule to the service engine 140. Also, as mentioned above, the context engines 110 on the hosts 200 in some embodiments resolve the policies based on collected contextual attributes, in order to generate rules for the service engines.

FIG. 12 illustrates an example of how the service engines 130 are managed in some embodiments. This figure illustrates multiple hosts 200 in a datacenter. As shown, each host includes several service engines 115, a context engine 110, a threat detector 132, a DPI module 135, several VMs 205, and an SFE 210. It also illustrates a set of controllers 1210 for managing the service engines 115, VMs 205, and SFEs 210. As mentioned above, the context engines 110 in some embodiments collects contextual attributes that are passed to the management servers in the controller set through a network 1250 (e.g., through a local area network, a wide area network, a network of networks (such as the Internet), etc.). The controller set provides a user interface for the administrators to define context-based service rules in terms of these collected contextual attributes, and communicates with the hosts through the network 1250 to provide these policies. The hosts also communicatively connected to each other through this network 1250.

Many of the above-described features and applications are implemented as software processes that are specified as a set of instructions recorded on a computer readable storage medium (also referred to as computer readable medium). When these instructions are executed by one or more processing unit(s) (e.g., one or more processors, cores of processors, or other processing units), they cause the processing unit(s) to perform the actions indicated in the instructions. Examples of computer readable media include, but are not limited to, CD-ROMs, flash drives, RAM chips, hard drives, EPROMs, etc. The computer readable media does not include carrier waves and electronic signals passing wirelessly or over wired connections.

In this specification, the term “software” is meant to include firmware residing in read-only memory or applications stored in magnetic storage, which can be read into memory for processing by a processor. Also, in some embodiments, multiple software inventions can be implemented as sub-parts of a larger program while remaining distinct software inventions. In some embodiments, multiple software inventions can also be implemented as separate programs. Finally, any combination of separate programs that together implement a software invention described here is within the scope of the invention. In some embodiments, the software programs, when installed to operate on one or more electronic systems, define one or more specific machine implementations that execute and perform the operations of the software programs.

FIG. 13 conceptually illustrates a computer system 1300 with which some embodiments of the invention are implemented. The computer system 1300 can be used to implement any of the above-described hosts, controllers, and managers. As such, it can be used to execute any of the above described processes. This computer system includes various types of non-transitory machine readable media and interfaces for various other types of machine readable media. Computer system 1300 includes a bus 1305, processing unit(s) 1310, a system memory 1325, a read-only memory 1330, a permanent storage device 1335, input devices 1340, and output devices 1345.

The bus 1305 collectively represents all system, peripheral, and chipset buses that communicatively connect the numerous internal devices of the computer system 1300. For instance, the bus 1305 communicatively connects the processing unit(s) 1310 with the read-only memory 1330, the system memory 1325, and the permanent storage device 1335.

From these various memory units, the processing unit(s) 1310 retrieve instructions to execute and data to process in order to execute the processes of the invention. The processing unit(s) may be a single processor or a multi-core processor in different embodiments. The read-only-memory (ROM) 1330 stores static data and instructions that are needed by the processing unit(s) 1310 and other modules of the computer system. The permanent storage device 1335, on the other hand, is a read-and-write memory device. This device is a non-volatile memory unit that stores instructions and data even when the computer system 1300 is off. Some embodiments of the invention use a mass-storage device (such as a magnetic or optical disk and its corresponding disk drive) as the permanent storage device 1335.

Other embodiments use a removable storage device (such as a floppy disk, flash drive, etc.) as the permanent storage device. Like the permanent storage device 1335, the system memory 1325 is a read-and-write memory device. However, unlike storage device 1335, the system memory is a volatile read-and-write memory, such a random access memory. The system memory stores some of the instructions and data that the processor needs at runtime. In some embodiments, the invention's processes are stored in the system memory 1325, the permanent storage device 1335, and/or the read-only memory 1330. From these various memory units, the processing unit(s) 1310 retrieve instructions to execute and data to process in order to execute the processes of some embodiments.

The bus 1305 also connects to the input and output devices 1340 and 1345. The input devices enable the user to communicate information and select commands to the computer system. The input devices 1340 include alphanumeric keyboards and pointing devices (also called “cursor control devices”). The output devices 1345 display images generated by the computer system. The output devices include printers and display devices, such as cathode ray tubes (CRT) or liquid crystal displays (LCD). Some embodiments include devices such as a touchscreen that function as both input and output devices.

Finally, as shown in FIG. 13, bus 1305 also couples computer system 1300 to a network 1365 through a network adapter (not shown). In this manner, the computer can be a part of a network of computers (such as a local area network (“LAN”), a wide area network (“WAN”), or an Intranet, or a network of networks, such as the Internet. Any or all components of computer system 1300 may be used in conjunction with the invention.

Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a machine-readable or computer-readable medium (alternatively referred to as computer-readable storage media, machine-readable media, or machine-readable storage media). Some examples of such computer-readable media include RAM, ROM, read-only compact discs (CD-ROM), recordable compact discs (CD-R), rewritable compact discs (CD-RW), read-only digital versatile discs (e.g., DVD-ROM, dual-layer DVD-ROM), a variety of recordable/rewritable DVDs (e.g., DVD-RAM, DVD-RW, DVD+RW, etc.), flash memory (e.g., SD cards, mini-SD cards, micro-SD cards, etc.), magnetic and/or solid state hard drives, read-only and recordable Blu-Ray® discs, ultra-density optical discs, any other optical or magnetic media, and floppy disks. The computer-readable media may store a computer program that is executable by at least one processing unit and includes sets of instructions for performing various operations. Examples of computer programs or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter.

While the above discussion primarily refers to microprocessor or multi-core processors that execute software, some embodiments are performed by one or more integrated circuits, such as application specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs). In some embodiments, such integrated circuits execute instructions that are stored on the circuit itself.

As used in this specification, the terms “computer”, “server”, “processor”, and “memory” all refer to electronic or other technological devices. These terms exclude people or groups of people. For the purposes of the specification, the terms display or displaying means displaying on an electronic device. As used in this specification, the terms “computer readable medium,” “computer readable media,” and “machine readable medium” are entirely restricted to tangible, physical objects that store information in a form that is readable by a computer. These terms exclude any wireless signals, wired download signals, and any other ephemeral or transitory signals.

While the invention has been described with reference to numerous specific details, one of ordinary skill in the art will recognize that the invention can be embodied in other specific forms without departing from the spirit of the invention. For instance, several figures conceptually illustrate processes. The specific operations of these processes may not be performed in the exact order shown and described. The specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. Furthermore, the process could be implemented using several sub-processes, or as part of a larger macro process. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims

Claims

1. A method for performing firewall operations on a host computer on which a plurality of machines execute along with a hypervisor that provides virtualization services, the method comprising:

at a firewall executing as a hypervisor service on the host computer outside of a virtual machine (VM),

receiving a data message of a data message flow (i) sent by the VM and (ii) forwarded to the firewall as the data message traverses along an egress path from the VM out of the host computer;

identifying a set of contextual attributes associated with the data message, the set of contextual attributes including a first identifier that specifies a type of traffic contained in the data message flow,

said identifying comprising (i) using a deep packet inspector executing on the host computer to perform deep packet inspection on the data message flow and to generate the first identifier specifying the traffic type contained in the data message flow and (ii) obtaining, from a context collector as a hypervisor service on the host computer, a second identifier that identifies an application that executes on the VM and that is the source of the data message flow;

using the first and second identifiers to identify a firewall rule to enforce; and

performing a firewall operation on the data message based on the identified firewall rule, said firewall operation resulting in the firewall (i) allowing the received data message to pass or (ii) dropping the received data message.

2. The method of claim 1, wherein the identified firewall rule specifies a firewall action, and performing the firewall operation comprises performing the firewall action of the identified firewall rule on the data message.

3. The method of claim 2, wherein the identified set of contextual attributes includes an identifier of a user, logged on to the VM.

4. The method of claim 3, wherein the user identifier is a group identifier.

5. The method of claim 3, wherein the user identifier is an identifier that identifies a single individual.

6. The method of claim 3, wherein the identified set of contextual attributes includes a threat level associated with an application that executes on the VM.

7. The method of claim 6, wherein the application is the source application that sent the data message.

8. The method of claim 1, wherein the set of contextual attributes includes at least one other contextual attribute that is not layer 2, layer 3 or layer 4 data-message header values.

9. The method of claim 1, wherein the contextual collector obtains the second identifier based on a set of contextual attributes received by the context collector from a guest introspector executing on the VM.

10. The method of claim 1, wherein using a deep packet inspector comprises obtaining the AppID from the context collector that executes as a hypervisor service on the host computer and that has the deep packet inspector generate the AppID before the firewall performs the firewall operation on the received data message.

11. The method of claim 1, wherein using a deep packet inspector comprises directing the deep packet inspector to examine a set of data messages of the flow to generate the AppID.

12. A non-transitory machine readable medium storing a firewall program for performing firewall operations for a first machine executing on a host computer with a plurality of other machines, the firewall program for execution on the host computer and comprising sets of instructions for:

receiving a data message of a data message flow sent by the first machine;

identifying a set of contextual attributes associated with the data message, the set of contextual attributes including a first identifier that specifies a type of traffic contained in the data message flow,

said identifying comprising using a deep packet inspector executing on the host computer to perform deep packet inspection on the data message flow and to generate the first identifier specifying the traffic type contained in the data message flow;

obtaining, from a context engine executing on the host computer, a second identifier that identifies an application that executes on the first machine and that is the source of the data message flow;

using the first and second identifiers to identify a firewall rule to enforce; and

performing a firewall operation on the data message based on the identified firewall rule, said firewall operation resulting in the firewall program (i) allowing the received data message to pass or (ii) dropping the received data message.

13. The non-transitory machine readable medium of claim 12, wherein the identified firewall rule specifies a firewall action, and performing the firewall operation comprises performing the firewall action of the identified firewall rule on the data message.

14. The non-transitory machine readable medium of claim 12, wherein the set of contextual attributes includes at least one other contextual attribute that is not layer 2, layer 3 or layer 4 data-message header values.

15. The non-transitory machine readable medium of claim 12, wherein the set of instructions for using a deep packet inspector comprises a set of instructions for obtaining the AppID from the context engine that executes on the host computer and that has the deep packet inspector generate the AppID before the firewall program performs the firewall operation on the received data message.

16. The non-transitory machine readable medium of claim 15, wherein the identified set of contextual attributes includes a threat level associated with an application that executes on the first machine.

17. The non-transitory machine readable medium of claim 16, wherein the application is the application that sent the data message.

18. The non-transitory machine readable medium of claim 16, wherein the application is not the application that sent the data message.

19. The non-transitory machine readable medium of claim 12, wherein the set of instructions for using a deep packet inspector comprises a set of instructions for directing the deep packet inspector to examine a set of data messages of the flow to generate the AppID.